Object permanence problems were administered to 19 rhesus monkeys as part of an ongoing investigation of representational thought in this species. On Invisible transfers, a treat was placed into one of three boxes, the box was moved next to another box, and the treat was transferred. The primates had to use representational thought (i.e., imagine the displacement of the object from one box to another) to solve the problem. On Visible transfers, space was left between the two boxes so that the transfer could be observed and on No transfer problems, the object remained in the first box. Thus, the latter two problems did not require representational thought. Contrary to previous results, the rhesus monkeys performed well on Invisible transfers. However, performance on Invisible and No transfers varied according to the boxes that were manipulated. Performance was high when adjacent boxes were manipulated and low when non-adjacent boxes were manipulated. This raises the possibility that the primates used a non-representational strategy to find the object. On adjacent trials, actions were carried with two boxes on one side of the board whereas no action was carried with the third box on the opposite side. Of the two boxes manipulated, one was always shown empty, thus leaving the correct box as the only alternative. However, on trials when non-adjacent boxes were involved, actions occurred across the board with the three boxes. One extreme box was passed in front of the middle box, reached the other extreme box and then was brought back. Consequently, the primates could not simply make a choice between two boxes. To determine whether the proposed strategy played a role in primate performance, we included probe trials in test sessions. On probe trials, the treat was placed in one of the extreme box. In contrast to regular adjacent trials, the experimenter then passed the baited box in front of the middle box and transferred (visibly or invisibly) or did not transfer the object from the other side of the middle box. As a result, the two boxes manipulated were not isolated on one side of the board. Preliminary results indicate significant impairment on probe problems, suggesting that rhesus monkeys can solve complex problems through the use of a non-representational strategy.